The goal of IDmeasurer package is to provide tools for
assessment and quantification of individual identity information in
animal signals. This package accompanies a research article by Linhart
et al.: ‘Measuring
individual identity information in animal signals: Overview and
performance of available identity metrics’, which can currently be
accessed at BioRxive.
The package is currently available at GitHub:
devtools::install_github('pygmy83/IDmeasurer', build = TRUE, build_opts = c("--no-resave-data", "--no-manual"))The package has been also submitted to CRAN and it should be soon
possible to install the released version of IDmeasurer from
CRAN with:
install.packages("IDmeasurer")This is a basic example which shows how to calculate individual
identity information in territorial calls of little owls
(ANspec example data):
library(IDmeasurer)Input data for the calculation of identity metrics in this package, in general, is a data frame with the first column containing individual identity codes (factor) and the other columns containing individuality traits (numeric).
summary(ANspec)
#> id dur df minf
#> 007a : 10 Min. :0.3680 Min. : 547.2 Min. : 476.6
#> 042a : 10 1st Qu.:0.5040 1st Qu.: 955.7 1st Qu.: 742.2
#> 045a : 10 Median :0.5680 Median :1014.0 Median : 820.3
#> 055a : 10 Mean :0.5733 Mean :1033.0 Mean : 798.7
#> 062a : 10 3rd Qu.:0.6320 3rd Qu.:1073.6 3rd Qu.: 890.6
#> 070p : 10 Max. :0.9760 Max. :1781.4 Max. :1101.6
#> (Other):270
#> maxf q25 q50 q75
#> Min. : 929.7 Min. : 570.3 Min. : 875.0 Min. : 898.4
#> 1st Qu.:1234.4 1st Qu.: 906.3 1st Qu.: 992.2 1st Qu.:1109.4
#> Median :1839.8 Median : 953.1 Median :1039.1 Median :1203.1
#> Mean :1609.0 Mean : 959.0 Mean :1049.6 Mean :1291.4
#> 3rd Qu.:1882.8 3rd Qu.:1007.8 3rd Qu.:1084.0 3rd Qu.:1523.4
#> Max. :1937.5 Max. :1203.1 Max. :1398.4 Max. :1750.0
#> This calculates HS metric for every single trait variable in the dataset.
calcHS(ANspec, sumHS=F)
#> vars Pr HS
#> 2 dur 0 1.13
#> 3 df 0 0.58
#> 4 minf 0 0.80
#> 5 maxf 0 1.06
#> 6 q25 0 1.04
#> 7 q50 0 1.48
#> 8 q75 0 0.93To calculate the HS for an entire signal, it is neccessary to have uncorrelated variables in dataset. Raw (correlated) trait variables need to be transformed into principal components by the Principal component analysis.
temp <- calcPCA(ANspec) Calculate HS for an entire signal.
calcHS(temp)
#> HS for significant vars HS for all vars
#> 4.68 4.68To see description of the example dataset, use:
?ANspecMore examples can be found in IDmeasurer vignette:
vignette('idmeasurer-workflow-examples')